Good practices in research

contributing to open science

We are here because we want to contribute to science. It’s a collective endeavor. We build on things that have been done in the past and we usually do that collaboratively. In very simplified terms, we collect data/build a model/develop a method, make analysis (data or model), produce some results, write a report on the analysis, publish (share with the community). This is the basis for the system to keep progressive. We add our little stone to the building. It needs to be solid for others to build on. The thing is that in the process, errors can be made. We can prevent that by making it easier for ourselves and others (collaborators, reviewers, users) to spot them. Also we want to be able to prevent fraud. And all of this is helps and improved by sharing our whole process, not only the final report but all the process to get there: this is the principle of open science which is a stepping stone in repoducable science. If things are open, I can reproduce them, I can check them and build on them.

Disclaimer

A paradigm shift

People used to think that it was ok not to disclose every step of our work, including data and code
~~An advantage~~

A paradigm shift

required by funding and journals (open science)
allows verification of the work (more rigorous, more reliable)
useful for your collaborators and yourself
increases visibility of the work (and citations)
fosters credibility and confidence in the scientific process
speeds up the scientific process

What does it mean?

Articles

A published article is just the tip of the iceberg that is the research process…

“An article about computational results is advertising, not scholarship. The actual scholarship is the full software environment, code and data, that produced the result.”

Claerbout and Karrenbach 1992.

The concepts

Wilson et al. 2016. Plos Comp Biol. https://arxiv.org/pdf/1609.00037.pdf

1. Project Organization : File organisation
2. Data management
3. Software : Code structure & analysis flow
4. Tracking Changes : version management
5. Manuscripts

1. Project Organization

“Organizing the digital artifacts of a project to ease discovery and understanding.”

= Research compendium

1. Project Organization

2. Data management

Saving both raw and intermediate forms
Documenting all steps: code data acquisition process when possible, any transformation of the data is coded (and the transformed data seen as a result)
Creating tidy data amenable to analysis

2. Data management: Tidy data

3. Software

Writing, organizing, and sharing scripts and programs used in an analysis.

Respect code conventions from your community, https://style.tidyverse.org/
Document your code
Document the workflow (what is done in what order with what)
Document software used (OS, versions of R, versions of functions)

4. Tracking Changes

Track the evolution of your project
Use a version control software (Git).

5. Manuscripts

Writing manuscripts in a way that leaves an audit trail and minimizes manual merging of conflict.

LaTeX (e.g. overleaf)

Markdown

Quarto (https://quarto.org)

5. Manuscripts

Reproducible research?

data

code

documentation (e.g. article)